J. Med. Chem. 1984,27, 726-733
726
was added to 600 mL of cold H20. The mixture was stirred vigorously as it was allowed to warm to room temperature. After 25 min, to the mixture was added 2.60 g (0.0310 mol) of NaHCO, (pH was still strongly acidic). After 1 h and 10 min, another 2.60 g (0.0310 mol) of NaHCO, was added (pH still was strongly acidic). After 1 h and 25 min, another 1.86 g (0.0221 mol) of NaHC0, was added, and the pH of the mixture was 4-5. After 1h and 40 min, the ether was removed under vacuum a t 30 “C. The remaining mixture was adjusted to pH 5-6 with 0.41 g (0.00488 mol) of NaHC03 and was extracted with 2 X 50 mL of CH2C12. The combined CH,Cl, extract was washed with 250 mL of water, dried over Na2S04,and filtered. The filtrate was concentrated to a pale yellow oil:, yield 11.0 g. The oil was chromatographed on silica gel in a Waters Prep 500 A apparatus with 3:2 hexanes/EtOAc: yield 7.58 g (69%) of a colorless oil 51, which solidified; mp 44-47 “C; IR (Nujol) 1745 (CO), 3428 and 3540 (NH) cm-’; NMR (CDC1,) 6 0.98 (2 t’s, 6 H, propyl CH3), 1.2-1.9 (m, 4 H, propyl C-2 CH,), 2.05 (s) and 2.12 (s) (9 H, acetyl CH,), 2.63 (2 t’s, 4 H, propyl C-1 CH,), 3.8-4.4 (m, 3 H, C-1 H and C-5 CH,), 4.84 (br s, 2 H, NH,), 4.9-5.2 (m, 1 H, C-4 H), 5.27 (dd, J = 7.5 and 2.5 Hz, 1H, C-2 H), 5.73 (dd, J = 7.5 and 2.5 Hz, 1H, C-3 HI; [aI2’~ +24.6” (c 1.05, CHCl,). Anal. (Cl~H3,NOsS2)C, H, N, S. 2,3-0-1sopropylidene-D-ribonol,4-lactone 54Di- tert -butyl phosphate) (49). The lactone 26 was treated with Ph3P in CBr4 to obtain the 5-bromo derivative. The bromo compound (1.2 g, 0.0047 mol) in 5 mL of dry DME was added dropwise to a refluxing solution of tetramethylammonium di-tert-butyl phosphate (1.3 g, 0.0047 mol) in 19 mL of dry DME under NP3 This mixture was refluxed for 1 h, cooled, and filtered. The filtrate was evaporated to an oil, which was dissolved in EtOAc and washed
with H20. The EtOAc solution was dried (Na2S04)and evaporated to a pale yellow oil 49: yield 1.7 g (94%); NMR (Me2SO-d8) 6 1.3-1.5 (several s, 24 H), 4.1 (dd, JpH = 6 Hz, Jd5 = 3 Hz, 2 H, H-5), 4.75 (s and m, 3 H, H-2, H-3, H-4). Isomerase Assay.24 A mixture containing 170 pL of 0.1 M histidine buffer a t pH 7.5, 30 p L of inhibitor solution, and 50 pL of enzyme preparation was preincubated for 2 min at 37 “C. The reaction was initiated by the addition of 50 pL of 10 mM Darabinose 5-phosphate or D-ribulose 5-phosphate and was terminated at 0,3,6,9, and 12 min by the addition of 50 pL of 1.5% cysteine hydrochloride solution, followed immediately by 1.5 mL of 25 N H,SOI. After being mixed, the samples were treated with 50 pL of 0.12% carbazole in 95% EtOH and incubated a t 37 “C for 30 min. The absorbance was read at 540 nm. The conversion of 1 pmol of D-A5F to 1 pmol of D-ribulose 5-phosphate gave a AOD of 8.2 & 0.34.
Acknowledgment. We thank B. S. Hurlbert, A. Ragouzeos, R. Crouch, and J. Miller for the NMR spectra; Dr. D. A. Brent, D. Rouse, R. L. Johnson, Jr., P. Chandrasurin, B. Soltman, J. Cichetti, and J. Sabatka for the mass spectra; P. Baker, N. T. Rodgers, S. Wrenn, J. Emery, T. Williams, 0. Webb, and C. A. Jenkins for technical assistance; and Drs. D. W. Henry, R. W. Morrison, and J. Burchall for consultation and support. Supplementary Material Available: NMR data for compounds 44-46,50, and 72-78 (2 pages). Ordering information is given on any current masthead page.
Oligonucleotide Structural Parameters That Influence Binding of to the 5’-0-Triphosphoadenylyl-(2’-*5’)-adenylyl-(2’~5’)-adenosine 5’-0 -Triphosphoadenyly 1- (2‘+5‘) -adenylyl-(2‘+5’) -adenosine Dependent Endoribonuclease: Chain Length, Phosphorylation State, and Heterocyclic Base Paul F. Torrence,*st Jiro Imai,t Krystyna Lesiak,t Jean-Claude Jamoulle,t and Hiroaki Sawait Laboratory of Chemistry, National Institute of Arthritis, Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20205, and Faculty of Pharmaceutical Sciences, University of Tokyo, Bunkyo-ku Tokyo, Japan. Received August 8, 1983 A number of 2’,5’-linked oligoadenylates and their analogues were prepared and evaluated for their ability to interact with the 5’-O-triphosphoadenylyl-(2’~5’)-adenylyl-(2’~5’)-adenosine (2-5A) dependent endoribonuclease of mouse L cells. The oligonucleotides were assayed for their ability to antagonize the action of 2-5A, to displace a radiolabeled probe from the 2-SA-dependent nuclease, or to inhibit translation in a cell-freesystem. These experiments demonstrated the following: (1)Three AMP residues in a 5’-phosphorylated oligonucleotide were needed for maximum interaction with the endonuclease, and higher oligomers (24 AMP residues) did not show significantly higher binding. (2) The third (2’-terminal) adenosine residue was required for optimal binding activity. (3) 5’-Phosphorylation of the oligonucleotide was necessary for maximum binding to the endonuclease, but the first (from the 5’ terminus) internucleotide phosphate of higher unphosphorylated or core oligomers, such as A2’p5’A2’p5’A2’p5’A, may partly replace the requirement for a 5‘-monophosphate moiety; in agreement with this, the 5’-methyl ester of 5‘pA2‘p5’A2‘p5‘A, i.e., Me-p5’A2’p5’A2’p5’A, was bound to the endonuclease as well as or better than the higher core oligomers but approximately 100 times more effectively than the trimer core, A2’p5’A2’p5’Ae (4) Base-modified analogues, such as P ~ ’ C ~ ’ P ~ ’ C ~ ’ P p5’U2’~5’U2’~5’U, ~’C, or p5’12’~5’12’p5’1,were at least 2000 times less effectively bound to the endonuclease than p5‘A2‘p5‘A2’p5‘A. (5) The triphosphate ppp5’12’p5’12’p5’1 was 10000 times less active than 2-5A as an inhibitor of translation. These latter two points implied the critical role of the adenine “-nitrogen and/or exocyclic amino group in the binding of 2-5A to the endonuclease.
replication by i n t e r f e r ~ n . Other ~ ~ ~ data have suggested The 2’,5’-phosphodiester bond l i k e d oligoriboadenylate 5’-O-triphosphoadenylyl-(2’~5’)-adenylyl-(2’~5’)- a possible role for the 2-5A system in the normal regulation adenosine (2-5A) is believed to be a mediator of interferon action (for reviews, see ref 1-4). Although it is probably (1) Lengyel, P. In “Interferon 3”; Gresser, I.; ed.; Academic Press: not involved in all the biological effects of interferon, there New York, 1980; pp 77-99. is strong evidence that 2-5A may be responsible for the (2) Revel, M. In “Interferon 1”;Gresser, I., ed.; Academic Press: New York, 1979; pp 101-163. inhibition of encephalomyocarditis virus and reovirus
(3) Baglioni, C. Cell 1979, 17,255-264. (4) Torrence, P. F. Mol. Aspects Med. 1982, 5, 129-171. (5) Williams, B. R. G.; Golgher, R. R.; Brown, R. E.; Gilbert, C. S.; Kerr, I. M. Nature (London) 1979, 282, 582-586.
NIADDK, NIH. University of Tokyo. 0022-2623/84/l827-0726$01.50/0
0 1984 American Chemical Society
2’,5’-Linked Oligoadenylates
of cellular differentiation or de~elopment.~-’~ Considerations such as the above suggest the use of 2-5A or its derivatives as potential antiviral or antitumor agents, moreover, 2-5A antagonist^'^ may find a niche in the treatment of interferon-induced However, the following two difficulties stand in the way of such applications:21p22(1)2-5A is not readily taken up by the intact eukaryotic cell, and (2) 2-5A is rapidly degraded by a phosphodiesterase(s) present in cell extracts. While chemical modification of the 2-5A structure might provide solutions to these difficulties, it is first necessary to define what positions of the 2-5A molecule are critical in its interaction with its target enzyme, RNase L or 2-5A-dependent endonuclease, which after activation degrades mRNA, resulting in inhibition of t r a n s l a t i ~ n . ~ ~In” ~this report, we examine the effects of oligonucleotide chain length, the extent of oligonucleotide phosphorylation, and the nature of the heterocyclic base on the ability of 2-5A analogues to interact with the 2-5A-dependent endonuclease.28ayb
(6) Nilsen, T. W.; Maroney, P. A.; Baglioni, C. J. Virology 1982, 42, 1034-1045. (7) Johnston, M. I.; Zoon, K. C.; Friedman, R. M.; De Clercq, E.; Torrence, P. F. Biochem. Biophys. Res. Commun. 1980, 97, 375-383. (8) Hovanessian, A. G.; Kerr, I. M. Eur. J . Biochem. 1978, 84, 149-159. (9) Etienne-Smekens, M.; Vassart, G.; Content, J.; Dumont, J. E. FEBS Lett. 1981,125,146-150. (10) Nilsen, T. W.; Wood, D. L.; Baglioni, C. J. Biol. Chem. 1981, 256, 10 751-10 754. (11) Stark, G. R.; Dower, W. J.; Schimke, R. T.; Brown, R. E.; Kerr, I. M. Nature (London) 1979,278, 471-473. (12) Krishnan, I.; Baglioni, C. Proc. Natl. Acad. Sci. U.S.A. 1980, 77,6506-6510. (13) Oikarinen, J. Biochem. Biophys. Res. Commun. 1982, 105, 876-881. (14) Besancon, F.; Bourgade, M. F.; Justesen, F.; Ferbus, D.; Thang, M. N. Biochem. Biophys. Res. Commun. 1981, 103, 16-24. (15) Kimchi, A. J. Interferon Res. 1981, 1, 559-569. (16) Kimchi, A.; Shure, H.; Revel, M. Eur. J. Biochem. 1981,114, 5-10. (17) Torrence, P. F.; Imai, J.; Johnson, M. I. h o c . Natl. Acad. Sci. U.S.A. 1981, 78,5993-5997. (18) Gresser, I.; Maury, C.; Tovey, M. G.; Morel-Maroger, L.; Pontillon, F. Nature (London) 1976,263,420-422. (19) Gresser, I.; Morel-Maroger, L.; Riviere, Y.; Guillon, J. C.; To-
vey, M. G.; Woodrow, D.; Sloper, J. E.; Moss, J. Ann. N.Y. Acad. Sci. 1980, 350, 12-20. (20) Gresser, I.; Tovey, M. G.; Maury, C.; Chouroulinkov,I. Nature (London) 1975,258, 76-78. (21) Torrence, P. F.; Imai, J.; Lesiak, K.; Johnston, M. I.; Jacobsen, H.; Friedman, R. M.; Sawai, H.; Safer, B. In “Interferons”; Merigan, T. C.; Friedman, R. M.; eds.; Academic Press: New York, 1982; pp 123-142. (22) Imai, J.; Johnston, M. I.; Torrence, P. F. J. Biol. Chem. 1982, 257, 12 739-12 745. (23) Clemens, M. J.; Williams, B. R. G. Cell 1978, 13, 565-572. (24) Baglioni, C.; Minks, M. A.; Maroney, P. A. Nature (London) 1978,273, 684-687. (25) Williams, B. R. G.; Ken, I. M.; Gilbert, C. S.; White, C. N.; Ball, L. A. Eur. J. Biochem. 1978,92, 455-462. (26) Slattery, E.; Ghosh, N.; Samanta, H.; Lengyel, P. Proc. Natl. Acad. Sei. U.S.A. 1979, 76, 4778-4782. (27) Schmidt. A.; Zilberstein, A.; Shulman, L.; Federman, P.; Berissi, H.; Revel, M. FEBS Lett. 1978, 95, 257-264.
Journal of Medicinal Chemistry, 1984, Vol. 27, No. 6 727
Results and Discussion Chemistry. The 2’,5’-linked base-modified oligonucleotides p5’12’~5’12’p5’1,p5’U2’p5’U2’p5’U,29avband the previously unreported p5’C2’p5’C2’p5’C were prepared by lead ion catalyzed polymerization of the 5’-phosphoroimidazolidate of inosine, uridine, and cytidine, respectively, in a similar manner to that used for the synthesis of 2’,5’-linked olig~adenylates.~~ The oligomers formed in each reaction were separated by QAE-Sephadex A25 column chromatography and then further purified by paper chromatography. Purity was established by both HPLC and TLC. The yields of the trinucleotides p5’12’~5’12’~5’1,p5’U2’p5’U2’p5’U, and p5’C2’p5’C2’p5’C were 4.5-6 % based on starting phosphoroimidazolidate. The oligoinosinate p5’12’p5’12’~5’1 was converted to ppp5’12’p5’12’p5’1 by the phosphoroimidazolidate method in the same way as already described for 2-5A.31 Thus, the trimer 5’-monophosphate was reacted with N,N’carbonyldiimidazole to form the corresponding 5‘phosphoroimidazolidate, which was reacted with pyrophosphate to give ppp5’12’p5’12’p5’1 and pp5’12’~5’12’~5’1 in 26 and 41% yield, respectively. Characteristic proton NMR signals (Table VII) and mobilities of the base-modified oligonucleotides on TLC (Table VI) were in agreement with the assigned structures. The assigned structures of the trinucleotides were further confirmed by enzyme digestion. Treatment of p5’12’~5’12’p5’1, p5’U2’p5’U2’p5’U, and p5’C2’p5’C2’p5’C with venom phosphodiesterase gave p5’1, p5’U, and p5’C, respectively, as the sole products. The above trinucleotides were, however, insensitive to RNase T2 or nuclease P1, which are specific for 3‘,5’ internucleotide bonds. Alkaline phosphatase digestion of p5’12’~5’12’~5’1,pp5’12’~5’12’~5’1, or ppp5’12’p5’12’p5’1 gave the core, 12’p5’12’p5’1, as the only product. When 12’p5’12’p5’1 was degraded by venom phosphodiesterase, it gave I and p5’I in a 1:2 ratio. Similarly, digestion of p5’C2’p5’C2’p5’C or p5’U2’p5’U2’p5’U with alkaline phosphatase gave the corresponding core oligomers, C2’p5’C2’p5’C or U2’p5’U2’p5’U. The route chosen for the preparation of the 5’methylphosphoryl derivative A2’p5‘A2’p5‘A, Le., Mep5’A2‘p5‘A2‘p5’A, was patterned after the earlier work of Schaller et and Cramer and N e ~ n h o f f e rwho , ~ ~ employed acid-catalyzed methanolysis to obtain the methyl ester of AMP. Since formation of 2’,3’-cyclic carbonate is sometimes typical of carbonyldiimidazole generation of an imidazolidate,34oxidation-reduction condensation35 was used to prepare the 5’-phosphoroimidazolidate of A2’p5’A2’p5’A9 Thus, the imidazolidate obtained as a (28) (a) Also referred to as RNase L or RNase F. (b) A portion of this work was presented at the UCLA Symposium on Molecular and Cellular Biology entitled “Chemistry and Biology of
Interferons: Relationship to Therapeutics”, Squaw Valley, CA, Mar 8-12, 1982:l and at the 5th International Round Table on Nucleosides, Nucleotides, and their Biological Applications, Research Triangle Park, NC, Oct 20-22, 1982. (29) (a) Sawai, H.; Ohno, M. Bull. Chem. SOC.Jpn. 1981, 54, 2759-2752. (b) Sawai, H.; Ohno, M. Chem. Pharm. Bull. 1981, 29, 2237-2245. (30) Sawai, H.; Shibata, T.; Ohno, M. Tetrahedron 1981, 37, 481-485. (31) Sawai, H.; Shibata, T.; Ohno, M. Tetrahedron Lett. 1979, 4573-4576. (32) Schalir, H.; Staab, H. A.; Cramer, F. Chem. Ber. 1961, 34, 1612-1633. (33) Cramer, F.; Neunhoeffer, H. Chem. Ber. 1962,45,1664-1668. (34) Hampton, A.; Kappler, F.; Picker, D. J. Med. Chem. 1982,25, 638-644. (35) Mukaiyama, T.;Hashimoto, M. Bull Chem. Soc. Jpn. 1971,44, 2284.
728 Journal of Medicinal Chemistry, 1984, Vol. 27, No. 6
Torrence et al.
Table I. Interaction of Oligonucleotides with 2-5A-Dependent Endonuclease as Determined by Antagonism of 2-5 Action and Radiobinding Assays no. 1 2 3
antagonism of 2-5A actiona
oligonucleotide
radio bin ding assay b
p5'A2'p5'A > 2 x 10-4 (>200)c 2 x 10-5 (40000) p5'A2'p5'A2'p5'A 1x (1) 8 X 10"O (1.6) p5'A2'p5'A2'p5'A2'p5'A 1x (1) 7 X lo-'' (1.4) 4 p5'A2'p5'A2'p5'A2'p5'A2'p5'A 5 x 10-7 (0.5) 9 X lo-'' (1.8) 5 p5'A2'p5'A2'p5'A2'p5'A2'p5'A2'p5'A 1 x 10-9 (2.0) 6 A2'p5'A 21 x 10-3 ( > i o o o ) c 7 A2'p5'A2'pS1A 21 x (2100) 8 X lo-' (1600) 8 A2' p5' A2'p5' A2'p 5' A 1 x 10-5 ( i o ) 2.2 X lo-' (40) 9 A2'p5'A2'p5'A2'p5'A2'pS A 7x (7) 1.3 X lo-' (26) 10 A2'p5'A2'p5'A2'p5'A2'p5'A2'p5'A 1.3 X lo-' (26) 11 Me-pCitA2'p5'A2'p5'A 2 x 10-9 (4) 4X (8000) 12 p5'A2'p5'A3'p 13 p 5 A2' p 5' A2' p 4 x 10-7 (800) p5A2'p5'A2'p5'A2'(3')p 3 x 10-7 (0.33) 5 x 10-'O (1) 14 15 p 5 C2' p5'C2'p5'C 21 x (>IOO)~,~ 3.3 X (6600) 16 p5'U2'pS'U2'p5'U x 10'~ ( > I O O ) ~ , ~ 1.1x (2200) 17 p 512' p 5' I2'p 5' I >1 x (>100)C~d 3.3 X (6600) 18 ppp5' I2'p 5'12'p 5' I >1.3 X (>130) 2.6 X 10.' (520) pppB'A2'p5'A2'pS' A 5x (1) - 19 a Molar concentration of oligonucleotide needed to give 50% prevention of 2-5A action. The value in parentheses refers t o the relative activity of the compound. Relative activity is defined as the relative amount o t achieve the half maximal effect. The higher the value, the less active was the oligomer. Molar concentration of oligonucleotide required to displace 50% of the radiolabeled probe from the endonuclease-nitrocellulose complex. Relative activity is given in parentheses (see footnote a ) . Compound had no effect at this, the highest concentration tested. At very high concentrations, these analogues caused some inhibition of translation. p5'C2'p5C2'p5'C gave 50% inhibition at 3 x lo-' M and p5'U2'p5'U2'p5'U gave 50% inhibition at 1 X M, while p5'12'p5'12'p5'1 required more than 1 x M to cause a 50% inhibition. This inhibition was not due to activation of the 2-5A-dependent endonuclease, since the inhibitions were not prevented by pStA2'p5'A2'p5'A (1.8 x M).
*
product of the reaction of p5'A2'p5'A2'p5'A with triphenylphosphine, imidazole, and dipyridyl disaide= gave, upon reaction with methanol under conditions of acid catalysis, Me-p5'A2'p5'A2'p5'A in good yield. In the same fashion, the 5'-methylphosphoryl derivative of the dimer, A2'p5'A, also could be prepared. Two separate approaches were employed for the syn= x I I / thesis of 2' (3') terminally phosphorylated oligonucleotides. For t h e synthesis of p5'A2'p5'A2'p5'A2'p, I p5'A2'p5'A2'p5'A2'p5'A was subjected to periodate oxidation and base-catalyzed elimination of adenine according to previously described m e t h o d ~ l o g y . ~Structure ~~~~~~~ confirmation was obtained by alkaline phosphatase digestion, which gave A2'p5'A2'p5'A as the sole product. M Oligonucleotide Different methodology was used in the preparation of the Figure 1. Comparison of ppp5/A2'p5'A2'p5'A (A) and dimer diphosphate pYA2'p5'A3'p. In this case, the linkage ppp5'12'p5'12'p5'1(0) as inhibitors of translation in extracts of isomermp5'A2'p5'A3'pS'A2'p5'A was digested with RNase mouse L cells programmed with encephalomyocarditisvirus RNA. T2. The other isomer, p5'A2'p5'A2'p, was prepared by The dashed line represents the level of translation in the absence periodate oxidation and base elimination of of any additions. p5'A2'p5'A2'p5'A. Both p5'A2'p5'A2'p and p5'A2'p5'A3'p were completely times slower than the hydrolysis of 3'-AMP by nuclease digested to A2'p5'A by bacterial alkaline phosphatase these results served to further confirm the assigned digestion for 3.5 h at 37 "C. The 2' (or 3') terminal structures. phosphate group was removed first, since TLC during the Biological Activities of t h e Oligonucleotides. Inearly part of the digestion revealed the presence of teraction of the various oligonucleotides with the 2-5Ap5'A2'p5'A. On the other hand, digestion of the above dependent endonuclease has been followed with three oligonucleotides with nuclease P1 gave differing results. different assays. The first involved determination of the Incubation of p5'A2'p5'A3'p with nuclease P1rapidly gave ability of the oligonucleotide to prevent the protein synthe expected product p5'A2'p5'A within 1h of incubation. thesis inhibitory effects of 2-5A in a mouse L-cell-free To the contrary, the other isomer, p5'A2'p5'A2'p, was reprotein synthesis system programmed with encephalosistant to the action of nuclease P1, showing no change myocarditis virus RNA.17 Since this antagonism of 2-5A after a 1-h incubation. Prolonged incubation (3.5 h) gave action is a result of competition of the oligonucleotide for a small yield (-10%) of p5'A2'p5'AV Since it has been the 2-5A binding site of the endonu~lease,3~**~ it provided previously established that hydrolysis of 2'-AMP is 3000 (38) Fujimoto, M.; Kuninaka, A.; Yoshino, H.Agr. Biol. Chern. -1
~~
~~
~~
(36) Neu, H. C.; Heppel, L. A. J. Biol. Chern. 1964,239,2927-2934. (37) Torrence, P. F.; Imai, J.;Johnston, M. I. Anal. Biochern. 1983, 129,103-110. Mukayama,T.; Hashimoto, M. Bull Chern. SOC. Jpn. 1971,44, 2284.
1974,38, 785-790. (39) Knight, M.; Cayley, P. J.; Silverman, R. H.; Wreschner,D. H.; Gilbert, C. S.; Brown, R. E.; Kerr, I. M. Nature (London) 1980, 288, 189-197.
2’,5’-Linked Oligoadenylates
a measure of the ability of the analogue to bind to the endonuclease. In this case, the concentration of analogue necessary to effect a 50% reversal of the inhibition of translation caused by 20 nM 2-5A was determined. The second assay revealed the ability of the oligonucleotide 5’-triphosphate to inhibit protein synthesis in the above cell-free system. In this instance, the concentration of oligomer necessary to cause half-maximal inhibition of translation was determined. The third method was the radiobinding assay developed by Knight et a1.39t40Here the concentration of oligomer needed to prevent 50% of the added ppp5’A2’p5’A2’p5’A2‘p5’A3’[32P]p5‘C3’p from binding to the endonuclease was established. The results of such studies are presented in Table I and Figure 1,and the following conclusions may be drawn from them. In Table I are given the concentrations of oligonucleotides necessary to achieve a 50% effect. Relative activity is defined as the relative amount required to give the 50% effect. The higher the value, the less active the oligomer. (1)When 5’-monophosphorylated and 5’-unphosphorphorylated oligonucleotides were considered, it was clear that for optimal antagonistic activity and endonuclease binding, a 5’4erminal monophosphate moiety was required. No unphosphorylated oligoadenylate (6-9) was as active as the 5’-monophosphorylated oligoadenylates (1-5) in antagonistic or radiobinding activity; i.e., removal of the 5’-terminal monophosphate moiety led to an approximate 10- to 100-fold increase in the concentration of oligomer required to prevent 2-5A action or to displace radiolabeled probe from the endonuclease. However, even though the trimer core (7)was approximately 100-fold less active as an antagonistic than its corresponding 5’-monophosphate (2) and also 1000 times less efficient than 2 in displacing the probe from endonuclease, the tetramer, pentamer, and hexamer cores (8-10) were only about 7-10 times less active than the tetramer or pentamer 5’-monophosphate (3 and 4) as 2-5A antagonists and only 15-30 times less efficient in displacing the probe. Thus, while 5’-unphosphorylated oligomers did not bind to the endonuclease as well as 5-phosphorylated oligomers, the differential was not so great in the tetramer and higher oligonucleotides as at the trimer level. Since elongation of the chain of the 5’-monophosphates (2-5, vide infra) beyond three nucleotides did not lead to substantial increases in antagonistic or radiobinding activity, it may be possible that the increased activity of the tetramer, pentamer, or hexamer core (8-10) as compared to the trimer core (7) may be rationalized by assuming that the first internucleotide residue (from the 5’ end) may “slip” onto the binding site normally occupied by the 5’-monophosphate group. Knight et a1.39940also have reported that the trimer core had substantially reduced binding activity when compared to the 5’-mono- or 5’-triphosphate. The behavior of the 5’-methyl ester of p5‘A2’p5’A2’p5’A, Le., 11, was consistent with this idea. Although, Mep5’A2’p5’A2’p5’A was bound to the endonuclease about 3 times less effectively than p5’A2’p5’A2’p5’A itself, it was about 400 times more active in binding than the trimer core, A2‘p5’A2’pYA. Nonetheless, for optimal endonuclease binding, the presence of a free 5’-terminal monophosphate seems to be a minimal requirement. (2) At least three adenosine residues were necessary for optimal antagonistic or endonuclease binding activity. Thus,p5’A2’p5’A (1) was more than 200 times less effective than higher oligomers (2-5) as antagonist of 2-5A action (40) Knight, M.;Wreschner, D. H.; Silverman, R. H.; Kerr, I. M. Methods Enzymol. 1981,79,216-227.
Journal of Medicinal Chemistry, 1984, Vol. 27,No. 6 729
and even less effective at displacing the probe from the endonuclease. The activity of the core oligoadenylates (6-10) also lend support to this conclusion of the “slippage” mechanism alluded to above. While the dimer and trimer “cores” (6 and 7) showed little or no evidence of endonuclease binding, the tetramer, pentamer, and hexamer cores (8-10) showed at least 10-fold higher activities, since the internucleotide linkage adjacent to the 5’-terminal adenosine may “slip” into the 5’-monophosphate binding site. The relative importance of the third (2’-terminal) residue of 2’,5’-(~A)~ (2-5A trimer triphosphate) can be further discerned by partly dissecting the terminal residue. When adenosine was removed from the tetramer 4, the resulting diphosphorylated oligonucleotide (14) became a more effective antagonist of 2-5A action and possessed the same apparent affinity for the endonuclease as did the tetramer (4) itself. On the contrary, a similar shortening of the trimer (3) to give the diphosphorylated dimer (13) produced a compound with dramatically reduced binding activity (Table I). The terminal isomer of 13, namely, 12, was bound even less effectively to the endonuclease, in accord with the observation41 that substitution of 3’,5’phosphodiester bonds for 2‘,5’-phosphodiester bonds in 2-5A results in a decrease of affinity for the endonuclease. Previous observation^^^ have established that 2’ (3’) terminal phosphorylation increases the resistance of the 25A-derived molecule to degradation. The increased antagonistic activity of 14 was in agreement with this. The results obtained with the 2’ (3’) terminally phosphorylated dimers (12 and 13) suggest that the third adenosine moiety may contribute substantially to endonuclease binding. That the ribose moiety of the terminal adenosine residue may not be critical in endonuclease binding was suggested by the data of Imai et who found that conversion of the 2’-terminal ribose moiety of the trimer p5‘A2’p5’A2‘p5’A to an N-hexylmorpholine derivative increased antagonistic capacity relative to the parent compound, thereby implying the lack of importance of the ribose ring of the 2’-terminal adenosine residue in binding to the endonuclease. However, Drocourt et al.& found that while compounds of the general formula ppp5’A2’p5’A2’p5’N (where N was any of the common nucleic acid bases) could not activate the endonuclease, they could antagonize 2-5A action as effectively as pA2’p5’A2’p5’A. This result implied that replacement of the 2’4erminal residue with other nucleosides did not decrease endonuclease binding. These data were obtained in a calcium phosphate coprecipitation system in which the concentrations of oligonucleotide actually entering the cell are impossible to ascertain. The significantly lower protein synthesis inhibitory activity of ppp5’A2‘p5’A as compared t o ppp5’A2‘p5’A2‘p5’A‘ is a reflection of the fact that three adenosine residues are required for maximum binding to the 2-5A-activated endonuclease. (3) As the oligonucleotide chain length was increased beyond three nucleotide residues no substantial increase in antagonistic activity or endonuclease binding occurred. The addition of single nucleotide units to the trimer (2) to give the tetramer (3) or pentamer (4) had only a marginal effect on activity. A similar situation obtains in (41) Lesiak, K.;Imai, J.; Floyd-Smith, G.; Torrence, P. F. J . Biol. Chem. 1983,258,13082-13088. Wreschner, D. H.; Gilbert, C. S.; Kerr, I. M. (42) Silverman, R.H.; Eur. J. Biochem. 1981,115,79-85. (43) Drocourt, J. L.;Dieffenbach, C. W.; Ts’o, P. 0. P.; Justesen, J.; Thang, M. N. Nucleic Acids Res. 1982,10, 2163-2174.
730 Journal of Medicinal Chemistry, 1984, Vol. 27, No. 6
comparison of the core tetramer, pentamer, and hexamer (8-10). These data agree with the earlier observation^^^ t h a t the various oligomers of 2-5A, i.e., ppp5’A2‘p5’A2‘p5‘A7 ppp5’A2’p5’A2’p5‘A2‘p5‘A, and ppp5‘A2’p5’A2’p5‘A2’p5‘A2‘p5’A, were equally effective as protein synthesis inhibitors. (4) All base-modified analogues of p5’A2’p5’A2’p5‘AA, including p5’C2’p5‘C2’p5‘C (E), p5’U2’p5’U2’p5’U (16), and p5’12’p5’12’~5’1 (17),showed dramatically reduced antagonistic and endonuclease binding activity. Apparently, just the simple elimination of the adenosine amino group was sufficient to lead to a >2000-fold loss of endonuclease binding affinity. Moreover, since the inosinate analogue (17)showed comparable activities to the cytidylate (15)and uridylate (16)analogues, it would appear that the adenine 6-amino group and possibly its N-1 nitrogen may be the greatest contributors to the binding of the adenine rings. Three reports have appeared concerning oligoinosinate analogues of 2-A.45-47 In these studies, the inosinate analogue was prepared by nitrous acid d e a m i n a t i ~ nor~ ~ by enzymatic synthesis by using the 2-5A synthetase4 from mouse L cells, although the former group of investigators found that ITP was not a substrate for the reticulocyte enzyme. In addition, other studiesem have indicated that nucleoside triphosphates other than ATP did not lead to homooligonucleotides but would incorporate into a 2‘,5’oligonucleotide chain only in the presence of ATP or an adenosine terminated trimer. In these latter cases, the nonadenosine NTP acts as a terminator of chain elongation. Finally, it may be noted that no structure confirmation of the putative oligoinosinates of the quoted have been reported. Since the synthetic p5’12’~5’12’p5’1 of this study was inefficiently bound to the 2-5A-dependent endonuclease, as judged by both antagonism and radiobinding assays, it was unlikely that the corresponding 5‘-triphosphate would be an effective endonuclease activator. In order to check the remote possibility that the 5-triphosphate moiety might be more critical in endonuclease binding in the case of the inosinate analogue than it is in the case of 2-5A, the triphosphate ppp5’12’p5’12’p5’1 was prepared by standard procedures. As is evident from Table I, the triphosphate of 2’,5’-oligoinosinate (18) was ineffective as an antagonist of 2-5A action and was more than 500 times less effective than 2-5A in displacing radiolabeled probe from the endonuclease. The experiments of Figure 1document that, in accord with the findings of Table I, ppp5’12’p5’12’p5’1 was a very poor inhibitor of protein synthesis, at least 10000 times less potent than 2-5A. Conclusions From this and related studies, it is possible to present the following tentative picture of the interaction of oligonucleotides with the 2-SA-activated endonuclease. Optimal binding of 5’-phosphorylated 2’,5’-oligoadenylates occurs Kerr. I. M.: Brown. R. E. Proc. Natl. Acad. Sci. U.S.A. 1978. 75, 256-260.
Suhadolnik, R. J.; Doetsch, D. W.; Reichenbach, N. L. Fed. Proc., Fed. Am. SOC. Exp. Biol. 1982, 41, 1455. Hughes, B. G.; Srivastava, P. C.; Muse, D. D.; Robins, R. K. Biochemistry 1983, 22, 2116-2126. Hughes, B. G.; Robins, R. K. Biochemistry 1983, 22, 2127-2135. Minks, M. A.; Benvin, S.; Baglioni, C. J. Biol. Chem. 1980,255, 5031-5035. Justesen, J.; Ferbus, D.; Thang, M. N. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 4618-4622.
Ferbus, D.; Justesen, J.; Besancon, F.; Thang, M. N. Biochem. Biophys. Res. Commun. 1981,100,847-856.
Torrence et al. Table 11. Preparation and Purification of 2’,5’-Oligoadenylate 5’-Monophosphates HP LC compd
fR,
yield,
min
%
TEAB gradient for DEAE-Sephadex colc
(PA), 19.7 29.5 0.5-0.30 M (0.27 M ) d 25.8 7.7 0.20-0.35 M (0.305 M ) d (PA), 0.25-0.40 M (0.34 M)d (PA), 30.6 2.2 0.30-0.45 M ( 0 . 3 7 M)d (PA), 33.3 0.85 0.35-0.50 M (0.41 M)d (PA), 35.1 0.55 Column: Zorbax ODS (21.2 mm X cm). Buffer A: 50 mM ammonium phosphate; buffer B: MeOH/H,O, 1:l. Flow rate: 9.0 mL/min. Gradient 0-20% B (in 40 min). Detector wavelength: 280 nm. Yields are calculated based on the are integration of each peak in the HPLC chromatogram. Column size: 1 . 6 X 8 cm; the column was eluted with a linear gradient (each 250 mL) of TEAB buffer, Concentration of TEAB buffer at which each oligoadenylate was eluted from the column.
when three or more AMP residues are present in the chain: longer oligomers do not show greater apparent affinities for the e n d o n u c l e a ~ e . ~ ~The , ~ ~2’-terminal *~~ adenosine residue of 2-5A may be required for maximum endonuclease binding, as well as the presence of 5’-terminal mono- or triphosphate groups, since nonphosphorylated or core oligomers show greatly reduced apparent affiniThis difference is, however, most dramatic at the trimer level. Tetramer and higher core oligomers show enhanced binding compared to the trimer core, albeit significantly less than the 5’-phosphorylated oligonucleotides (Table I). This raises the possibility that one of the internucleotide phosphates may replace, albeit less effectively, the 5’-terminal phosphate group in binding to the endonuclease. Interaction of 2-5A with the endonuclease also occurs through the adenine rings; moreover, the adenine N’ nitrogen and/or adenine 6-amino group probably plays a critical role in the binding interaction (Table I and Figure 1). This conclusion is also supported by the o b ~ e r v a t i o nthat ~ ~ the 1,iWethenoadenosine analogue of 2-5A binds very poorly to the 2-SA-dependent endonuclease. Endonuclease binding also involves the nature of the internucleotide bond, since replacement of one or the other 2’,5’-linkages of 2-5A by 3’3‘ linkages gives isomers that bind to the endonuclease 20-50 times less effectively than 2-5A itself; moreover, 3-5A, i.e., ppp5’A3’p5’A3’p5’A7 is 10000 times less effective than 2-5A in endonuclease binding or as an inhibitor of translation?1 Finally, interaction with the 3’-hydroxyl groups of 2-5A is not important for oligonucleotide binding to the endonuclease but is vital for endonuclease activation, since analogues in which all of the 3‘-hydroxyl groups have been replaced by methoxy or hydrogen bind well to the 2-5Adependent endonuclease (albeit less effectively than 2-5A) but do not activate it to degrade RNA or to inhibit protein synthesi~.~~-~~ Experimental Section High-performanceliquid chromatography was carried out with either a Hitachi 638 apparatus using an RPC-5 column (4 mm x 25 cm) and elution with a linear gradient of NaClO, (51) Williams, B. R. G.; Kerr, I. M. Nature (London) 1978, 276, 88-90. (52) Lesiak, K.; Torrence, P. F. FEBS Lett. 1983, 151, 291-296. (53) Baglioni, C.; D’Alessandro, S. B.; Nilsen, T. W.; de Hartog, J. A. J.; Crea, R.; Van Boom, J. H. J . Biol. Chem. 1981, 256, 3253-3257. (54) Sawai, H.; Imai, J.; Lesiak, K.; Johnston, M. I.; Torrence, P. F. J.Biol. Chem. 1983, 258, 1671-1677. (55) Haugh, M. C.; Cayley, P. J.; Serafinowska, H. T.; Norman, D. G.;Reese, C. G.; Kerr, I. M. Eur. J.Biochem. 1983,132, 77-84.
2’,5‘-Linked Oligoadenylates
Journal of Medicinal Chemistry, 1984, Vol. 27, No. 6 731
Table 111. Characteristic Proton NMR Signals of 2’ ,5’-Oligoadenylate 5‘-Monophosphates chemical shift,a 6 compd
Ar ring protons (C-2 or C-8)
anomeric protons (C-1’)e
8.08 (1H), 7.93 (1 H ) , 7.83 (1 H), 7.75 (1H ) 7.98 (1H), 7.97 (1H ) , 7.80 (1H), 7.76 (2 H ) , 7.73 (1 H)
5.97 (1H, d , J = 3.0 Hz), 5.64 (1H, br s) (PA), 5.89 (1H, d , J = 3.0 Hz), 5.75 (1H, d , J = 3.7 Hz), (PA), 5.64 (1H, d , J = 4.0 Hz) (PA), 7.96 (1H), 7.93 (1H), 7.70 ( 3 H), 7.65 (1H), 7.62 (1H), 5.85 (1H, d , J = 2.7 Hz), 5.68 (1H, d , J = 3.6 Hz), 7.61 (1H) 5.60 ( 2 H, m) (PA), 7.94 (1H), 7.91 (1 H), 7.70 (1H), 7.68 (1H), 7.67 (1H), 5.84 (1H, d , J = 3.0Hz), 5.67 (1H, d , J = 4.1 Hz), 7.64(1H),7.63(1H),7.61(1H),7.59(1H),7.53 5.56 ( 3 H, m) (1 H) (PA), 7.94 (1 H), 7.93 (1H), 7.72 (1H), 7.71 ( 2 H), 7.69 (1H), 5.90 (1 H, d , J = 4.0 Hz), 5.69 (113, d , J = 3.8 Hz), 7.68 ( 2 H), 7.67 (1H), 7.65 (1H), 7.56 (1H), 7.54 5.60 (4 H, m) (1 H) a Chemical shifts are determined in D,O, with acetone ( 6 2.05) as an internal standard. All aromatic protons appeared as singlets. Doublets and multiplets are indicated as d and m. (0.0-0.05 M) buffered with Tris-acetate (pH 7.5) or with a Table IV. Characteristics of Synthetic Methylphosphoryl Beckman instrument with Model llOA pumps with columns and Oligoadenylates on HPLC eluents as indicated in the text. HPLC t ~rnin, ~ Paper chromatography was performed with Whatman 3MM compd Accupak (anal.) Zorbax (prep) paper by the descending technique and with concentrated 1propanoljconcentrated NH40H/Hz0 (55:1035) as developing 7.25, 7.26b 10.22, 10.26b (PA), solvent. 11.49 Me(p-41, 9.01 Thin-layer chromatography (TLC) was performed on 7.28 10.58 (PA), PEI-cellulose F in 0.25 M NH4HC03(solvent l),on cellulose F 9.95 11.83 Me(PA 1 3 in 1-propanol/concentrated NH40H/Hz0 (55:1035) (solvent 2), a HPLC experiments were carried o u t on a Beckman or on cellulose F in saturated ammonium sulfate/O.l M sodium instrument with Model 110 A pumps with detector at 280 acetate/2-propanol (79:19:2) (solvent 3). nm, with an Accupak column (flow rate 1 mL/min), and Materials. Inosine 5’-monophosphate (Na salt), uridine 5’with a Zorbax column (flow rate 2.5 mL/min). Both colmonophosphate, and cytidine 5’-monophosphate were from Yaumns were eluted with a linear gradient of 50 mM masa Co. All 2’-5’ linked core oligoadenylates, except A2’p5’A NH,H,PO, (pH 7.0) (A) mixed up t o 70% of methanol/ and A2’p5’A2’p5’A, were from Calbiochem. (La Jolla, CA). Sigma water (1:1)(€3) in 10 min. Elution was then continued Chemical Co. (St. Louis, MO) was the source for A2’p5’A, A3’p5‘A, t o 100%(B) in 5 min. Elution t o 100% in 10 min. and A3’p5‘A3’p5‘A. TLC plates were from Merck. DEAEAfter 10 min, the flow rate was reduced t o 0.75 mL/ Sephadex and QAE-Sephadex were from Pharmacia. Enzymes min . such as bacterial alkaline phosphatase or venom phosphodiesterase were from Worthington Biochemical Corp. Nuclease P1was digestion and HPLC purification was employed instead of direct obtained either from Sigma or Yamasa. All other materials obapplication of the mixture to the QAE-Sephadex column chrotained commercially were either reagent or HPLC grade. 2-5A matograph~.~~ and 2-5A trimer core were prepared as described ea~-lier.~~v~’ The polymerized reaction mixture (75 000 ODzso)was first Imidazole was purchased from J. T. Baker Chemical Co. Tritreated with Chelex 100 (100-200 mesh, NH4+form, 30 mL) at phenylphcephine,dipyridyl disulfide, and AMP were from Aldrich. room temperature for 1h to remove Pb2+ion. After filtration NMR Spectra. The 400-MHz proton NMR spectra were of Chelex, the filtrate was evaporated under vacuum, and ethanol recorded with a JEOL GX-400 instrument at 27 “C. The 220-MHz (40 mL) was added to the dried residue. The insoluble precipitate proton NMR spectra were obtained on a Varian HR 220 specwas collected by filtration and redissolved in HzO (40 mL) for trometer. In some instances, samples first were purified by paper the absorbance measurement (68000 OD,). To this solution were chromatography on Whatman 3MM paper and then passed added 100 mM morpholine ethane sulfonic acid buffer (pH 6.0, through a short column of Dowex 50W X 8 (Na+ form). The 5.0 mL), 1mM ZnClz (5.5 mL), and RNase P1 (1.0 mg, 800 units). solution was evaporated in vacuo, and then DzO was added and The reaction mixture was incubated at 40 OC for 2.5 h, followed evaporated (3 X 0.2 mL). Samples were generally made 10-20 by heating at 100 OC for 10 rnin and deproteinization with a mM in DzO and in some cases were buffered with sodium mixture of CHCl,/isoamyl alcohol (5:2, 10.0 mL). The aqueous phosphate (pH 6.6). Chemical shifts were recorded relative to layer was separated and concentrated under vacuum to approxan external TSP standard. imately 10 mL. The crude mixture of 2‘,5’-(pA), ( n = 2-6) was The preparation of mouse L cell extracts, encephalomyocarditis then separated by reverse-phase HPLC (Zorbax ODs column, 2.12 virus RNA, and the techniques and conditions for cell-free protein X 25 cm). Approximately 4000 ODzwuniwof the crude mixture synthesis and antagonism of 2-5A action assays have been dewas injected, and the column was eluted with a linear gradient scribed elsewhere.17*s8 Radiobinding assays were performed of 0-20% buffer B in 40 min (buffer A: 50 mM ammonium according to Knight et a1.,39,40 w i t h phosphate, pH 7.0; buffer B: 1:l methanol/HzO, flow rate 9.0 ppp5‘A2‘p5’A2’p5’A2‘p5‘A3’[32P]p5‘C3‘p of specific activity 3000 mL/min). Corresponding peaks for 2’,5’-(pA), ( n = 2-6) were Ci/mmol (Amersham, Chicago, IL). This assay measures the collected separately. In these separation conditions, all desirable capacity of an oligonucleotide to displace the radiolabeled probe 2’,5’-linked oligoadenylates were eluted from the column before from the endonuclease as determined by a nitrocellulose filter 25 min so that the all other byproducts that stayed in the column binding assay. longer could be separated from them. The column was then Preparation of 2’,5’-Oligoadenylate5’-Monophosphates washed with 100% buffer B to remove other products. The whole [2’,5’-(pA),,n = 2-61. 2’,5’-Oligoadenylate monophosphates (PA), HPLC operation was repeated several times until the substantial were prepared by lead ion catalyzed polymerization of adenosine amount of higher oligoadenylates were pooled. The collected 5’-phosphoimidazolidate(ImpA) basically according to the method 2’,5’-(pA), ( n = 2-6) eluates were applied individually to a short described by Sawai et al.30 For the purification of each oligoDEAE-Sephadex column (1.6 cm X 8 cm) to remove phosphate adenylate from the crude reaction mixture, however, RNase P1 buffer. The retention time, the triethylammonium bicarbonate (TEAB) buffer gradient condition on the DEAE-Sephadex col(56) Imai, J.; Torrence, P. F. J. Org. Chem. 1981, 46, 4015-4021. umn, and the yield for each oligomer are listed in Table 11. The (57) Imai, J.; Torrence, P. F. Methods Enzymol. 1981, 79, 233-244. proper fractions from the DEAE-Sephadexcolumn were combined (58) Torrence, P. F.; Friedman, R. M. J. Biol. Chem. 1979. 254. and concentrated under vacuum. The resulting residue was 1259-1267. coevaporated with HzO several times until the TEAB buffer was
732 Journal of Medicinal Chemistry,1984, Vol. 27, No. 6
Torrence et al.
Table V. Characteristic Proton NMR Signals of 5’ -0-(Methylphosphoryl)2’ ,5’-Oligoadenylates chemical shifts,a s compd
Ar ring protons (C-2 or C-8) 8.19 (1H), 7.93 (1H) 8.02 (1 H), 7.92 (1HI, 7.88 (1 H), 7.76 (1H) 7.98 (1 H), 7.90 (1H), 7.85 (1H), 7.76 (1 H), 7.73 (1 H), 7.65 (1H)
anomeric protons ( C - l ’ ) c 5.87 (1H, d, J = 5.0 Hz) Me(pA)d 6.13 (1H, d , J = 2.5 Hz), 6.06 (1H, d, Me(pA), J = 3.0 Hz) 5.88 (1H, br s), 5.76 (1H, d , J = 5.0 Hz), Me(pA), 5.60 (1 H, d , J = 4.0 Hz) a Chemical shifts were determined in D,O with acetone ( 6 2.05 ppm) as an internal standard. appeared as singlets. Doublets are indicated as d. Calcium salt. completely removed. A small amount of each 2’,5’-oligoadenylate thus prepared was reinjected into the HPLC column to assure the purity. All compounds [2’,5’-(pA),, n = 2-61 prepared above showed resistance to RNase PI and RNase Tzbut were degraded completely to 5’-AMP with snake venom phosphodiesterase. lH NMR data for these compounds are presented in Table 111. Preparation of 5’-0 -(Methy1phospho)adenylyl-(2’-5’)adenylyl-(2’+5’)-adenosine (Me-p5’A2’p5’A2’p5’) and 5’0 -(Methylphosp hory1)adenylyl-(2’+5’)-adenosine (Mep5’A2’p5’A). According to the preparation of adenosine 5’phosphoroimidazolidate described by Mukaiyama and Hashimot0,3~p5‘A2’p5’A2’p5’A (1080 Azss units, 29.21 pmol, triethylammonium salt) was dissolved in dry MezSO (150 pL). Imidazole (150 pmol), triphenylphosphine (60 pmol, 15.75 mg), freshly redistilled triethylamine (5 pL), and dipyridyl disulfide (60 pmol, 13.2 mg) were added to the initial solution. The yellow mixture was stirred at room temperature for 40 min and then transferred dropwise and under stirring to a 10-mL solution of sodium iodide in dry acetone (60 pmol/mL). The resulting precipitate was centrifuged and washed three t i e s with dry acetone (each 5 mL), and the white residue was dried over P205in vacuo for 3 h. The imidazolidate (916 A258 units, 24.7 pmol) was added to 500 WLof anhydrous methanol, and the mixture was treated at 4 “C with 47.8 pmol(5 mg) of imidazole hydrochloride. The reaction mixture was kept stirring for 3 days at 4 “C. The excess methanol was then removed under vacuum. The residue was dissolved in Me2S0, and the crude sodium salt was collected as described above. The precipitated salt was dissolved in water (1mL) and purified by HPLC (reverse phase, Zorbax ODS column, 0.94 X 25 cm). The column was eluted with two stepwise linear gradients of 50 mM NH4H2P04(pH 7.0) (A) mixed up to 70% of methanol/water (1:l)(B) in 10 min and then to 100% (B) in 10 min. The methyl ester was collected at a retention time of 11.85 min. It’s concentration in the reaction mixture was 82% according to the integration of the area of each peak in the HPLC chromatogram. The collected fractions containing the 5’-methyl ester were pooled, evaporated under vacuum, and applied to a DEAE-Sephadex A 25 (HC03-) column (20 X 1.0 cm), in order to remove ammonium phosphate buffer. The column was eluted with a linear gradient (250/250 mL) of 0.1 to 0.3 M TEAB buffer (pH 7.6). Fractions from 21 to 35 contained the 5’-O-methyl ester of the trimer monophosphate. They were pooled and evaporated repeatedly with water to remove TEAB buffer. The dry residue was dissolved in anhydrous methipol, and the sodium salt was isolated in the same manner as described above. The yield was 68.5% (740 Az& units based on monophosphate). A similar procedure was employed to prepare the 5’-0(methylphospho)adenylyl-(2’~5’)-adenosine.Thus, starting from 850 AZb8units (32.44 pmol) of the dimer monophosphate, esterification and HPLC purification were carried out as previously described. The elution gradient was from 0.05 to 0.15 M TEAB; fractions from 52 to 70 were collected and contained 450 Am units (17.18 pmol). The yield was 55% based on the monophosphate. HPLC characteristics and pertinent NMR data on these compounds are presented in Tables IV and V. Synthesis of the Base-Modified2’,5’-LinkedTrinucleotides p5’12’p5/12’p5’1, p5’C2’p5’C2’p5’C, and p5’U2’~5’U2’~5’U. 2’,5’-Linked triinosinate (p5’12’~5’12’~5’1)and triuridylate (p5’U2’p5’U2’p5’’U) were prepared by the published proced u r e ~ . The ~ ~ synthesis ,~~ of the tricytidylate (p5’C2’p5’C2’p5’C) was by a method similar to that used for the preparation of 2’,5’-linked oligoadenylates (Sawai et al.). Cytidine 5’-mono-
0 I1
CH,O-P-0 3.38 ( 3 H, d, J = 9.5 Hz) 3.03 ( 3 H, d , J = 9.5 Hz) 2.99 ( 3 H, d , J = 9.5 Hz) All aromatic protons
Table VI. TLC Mobilities of the Base-Modified Trinucleotides compd
Ria
Rfb
RfJC
PC pC2‘p5’C2‘p5’C c 2‘p5‘c2‘p5’c
0.70 0.48 0.74
0.42 0.21 0.47
0.79 0.70 0.67
PU pu2’p5‘u2’p5’U U2’p5’ U2’p 5’u
0.60 0.11 0.14
0.40 0.28 0.45
0.80 0.69 0.61
0.52 0.41 PI pI2’p5’12’p5’1 0.10 0.31 I2’p5‘12’p 5’I 0.14 0.48 PPI 2’p5’12‘p5‘I 0.08 0.24 pppI2’~5’12’p51 0.04 0.22 a Solvent system 1. Solvent system 2. system 3.
0.73 0.69 0.56 0.76 0.82 Solvent
phosphate (1 mmol) was converted to cytidine 5’-phosphoroimidazolidate (ImpC) by reaction with imidazole, triphenylphosphine, and 2,2’-dipyridyl disulfides and was obtained as the sodium salt in a yield of >go%. The polymerization of ImpC was accomplished by using lead nitrate as a catalyst. The reaction mixture (8 mL) contained 0.4 mmol of ImpC and 0.1 mmol of lead nitrate in 0.2 M imidazole buffer (pH 7.0). The mixture was maintained at room temperature for 5 days with constant stirring. Treatment of the reaction mixture with EDTA (0.5 mL of 0.25 M) gave a homogeneous mixture, indicating formation of the lead-EDTA complex. The resultant solution was applied to a column (30 mm X 45 cm) of QAE-Sephadex A25, and the produds were eluted with a stepwise linear gradient of triethylammonium bicarbonate buffer. The fractions containing p5’C2’p5’C2’p5’C were collected and evaporated in vacuo at